OBJECTIVES: To determine fall risk factors in nursing home and intermediate-care residents who can and cannot stand unaided. DESIGN: Prospective cohort study. SETTING: Residential elderly care facilities in Sydney, Australia. PARTICIPANTS: One thousand people aged 65 to 103 (mean age + standard deviation: 85.0 + 7.4). MEASUREMENTS: Accidental falls. RESULTS: Fall rates were highest in those with fair standing balance, intermediate in those with the best standing balance, and lowest in those with the worst standing balance. This nonlinear pattern was even more striking when subjects were categorized according to their standing balance and ability to rise from a chair. Using this dual classification, fall rates were highest in those who could rise from a chair but could not stand unaided (81%) and lowest in those who could neither rise from a chair nor stand unaided (48%). In residents who could stand unaided, risk factors included increased age, male sex, higher care classifications, incontinence, psychoactive medication use, previous falls, and slow reaction times. In contrast, quite different risk factors were evident in residents who could not stand unaided, with a number of known fall risk factors (previous stroke, reduced ability to rise from a chair, slow reaction times) being associated with fewer falls. In this group, risk factors were intermediate versus nursing home care, poor health status, psychoactive medication use, Parkinson's disease, previous falls, and being able to get out of a chair. CONCLUSION: The findings indicate that there are different risk factors for falls for people living in residential aged care facilities who can and cannot stand unaided. These findings provide important information for developing fall-prevention strategies and suggest that those who can stand unaided but have multiple falls risk factors constitute the highest priority group for such interventions.